Usage Audits and True-Ups: Reconcile the Contractual Meter

What is a licence true-up?

A true-up reconciles covered entitlement with usage under the reviewed terms. It can produce a charge, credit or no adjustment. Ask whether it measures peak access, active users, named identities or successful requests, and whether any change applies retrospectively or prospectively. The label does not define the rule.

Distinguish an invoice correction from a rights problem. Paying for an extra seat does not necessarily authorize an unapproved affiliate or an excluded form of redistribution. Conversely, an erroneous log classification should not become evidence of unauthorized use merely because it increases a draft invoice. Keep permission, measurement and financial settlement as separate questions so each can be resolved by the right reviewer.

Define the observation population before auditing

Specify the reporting period, timezone, identity mapping and treatment of test accounts, deactivated users and service credentials. If one person uses several identifiers, the agreement may or may not treat them as one seat. Record that interpretation explicitly. A technical system that counts tokens cannot independently determine a people-based charging rule, and deleting duplicate log events does not establish that two distinct users should share an entitlement.

WIPO's licensing checklist asks whether royalty reports and inspections are appropriate to verifying the agreed payment basis. Apply that bounded purpose to model access: request the evidence needed to reconstruct the meter, not unrestricted access to the buyer's investment files. Agree notice, reviewer confidentiality, inspection method and who bears costs. Those are negotiation questions rather than universal rights conferred by the existence of a subscription.

Reperform a synthetic seat-month reconciliation

Assume an invented agreement includes eight named seats each month, with additional seat-months priced at an illustrative USD 60. The accepted counts for three successive months are nine, eleven and ten. Additional seat-months are one plus three plus two, totaling six, and the proposed adjustment is USD 360. These are synthetic contract assumptions, not supplier pricing, standard industry terms or an assertion that peak-seat billing is appropriate.

Now suppose one identity in the middle month is a disputed test account. Removing it, if the reviewed terms support exclusion, reduces that month's additional count to two. The adjustment becomes five seat-months times USD 60, or USD 300. The USD 60 difference is a classification question with an arithmetic consequence. Present both versions until the classification is resolved instead of averaging them into an apparently agreed bill.

Constrain evidence handling and privacy exposure

Usage evidence can reveal employee identities, research interests and commercially sensitive schedules. Ask whether pseudonymous identifiers or a controlled reconciliation can answer the audit question without exporting raw activity records. Pseudonymization is not the same as anonymization, particularly when the reviewer can reconnect identifiers to people. Define secure transfer, retention and disposal for the audit packet rather than treating the inspection clause as permission for unlimited reuse.

The ICO's data-minimisation guidance explains the UK GDPR principle for personal information within its scope. It does not impose that framework on every buyer worldwide. Its relevance here is a precise review question: are the requested personal details adequate and necessary for the stated purpose? The actual privacy basis and responsibilities require jurisdiction-specific assessment alongside the commercial audit process, not after sensitive logs have already been distributed.

Specify disagreement and correction mechanics

Identify how the buyer challenges a count, how evidence is exchanged and which agreement version supplies the rate. Ask whether a credit, minimum commitment or earlier payment has already covered the disputed usage. Otherwise a mathematically correct adjustment may charge the same entitlement twice. The record should show the original invoice, proposed revision and settlement status without silently overwriting the evidence that created the disagreement.

Avoid relying on an undefined materiality threshold or an assumed right to suspend access during any dispute. Those consequences belong in the reviewed terms. A practical decision asks whether the buyer can continue its permitted workflow while a genuine counting issue is examined and whether a supplier can obtain necessary evidence promptly. Neither question can be answered solely by adding more rows to a cost comparison.

Use findings to improve the next period

Correct the cause as well as the invoice: stale identities, undocumented sharing or an unimplementable definition can recur. Preserve old rules for old periods and date any reviewed change. Do not silently backfill a later interpretation into historical reporting.

At renewal, ask whether the meter matches demand and whether evidence collection is proportionate. Repeated service-account disputes may need clearer terms. Counting retries twice or applying a later rate to an earlier period can distort charges. A clean reconciliation establishes no predictive merit.

Source notes

1. WIPO GREEN: Licensing Checklist, royalty reporting and inspections
2. ICO: Principle (c), Data minimisation, public UK guidance